# Claude Fable 5.2 vs GPT-6 Astra

As of Nov 22, 2026, Claude Fable 5.2 leads GPT-6 Astra in 7 of 9 shared categories: General LLM, Reasoning, Coding, Coding › Web dev, Long context, Factuality, and Legal.

## Which is better for coding?

As of Nov 22, 2026, Claude Fable 5.2 leads GPT-6 Astra in Coding at #2 against #3.

## Pricing

Claude Fable 5.2: $16 per 1M blended tokens.

GPT-6 Astra: $20 per 1M blended tokens.

Claude Fable 5.2: closed weights.

GPT-6 Astra: closed weights.

## Benchmarks side by side

| Benchmark | Claude Fable 5.2 | GPT-6 Astra |
| --- | ---: | ---: |
| ARC-AGI | 98.5% | 98.5% |
| Code Migration | 83.3 | 80.6 |
| Excel Modeling Benchmark (EMB) | 77.4 | 72.7 |
| Finance Agent | 69.5 | 60.4 |
| GDP.pdf | 51.8% | 34.2% |
| GPQA Diamond | 93.9% | 95.8% |
| Legal Research Bench | 46.2 | 39.4 |
| LMArena Text (style-controlled, coding) | 1526 | 1548 |
| LMArena Text (style-controlled, hard prompts) | 1517 | 1501 |
| LMArena Text (style-controlled, longer query) | 1518 | 1485 |
| LMArena Text (style-controlled, overall) | 1503 | 1475 |
| LMArena Text factuality | 1483 | 1468 |
| LMArena Vision (style-controlled) | 1305 | 1281 |
| LMArena WebDev | 1824 | 1764 |
| MysteryMechanism | 51.8 | 53.2 |
| ProofBench | 100 | 99.0 |
| SimpleQA Verified | 94.0% | 75.6% |
| Tax Agent Bench | 32.1 | 20.7 |
| Vals Index | 71.3 | 65.0 |
| Vibe Code Bench | 93.8 | 89.6 |

## Verdict

As of Nov 22, 2026, Claude Fable 5.2 leads GPT-6 Astra in 7 of 9 shared categories: General LLM, Reasoning, Coding, Coding › Web dev, Long context, Factuality, and Legal.

- Which to pick, General LLM: As of Nov 22, 2026, Claude Fable 5.2 leads GPT-6 Astra in General LLM at #5 against #30.
- Which to pick, Reasoning: As of Nov 22, 2026, Claude Fable 5.2 leads GPT-6 Astra in Reasoning at #10 against #31.
- Which to pick, Science: As of Nov 22, 2026, GPT-6 Astra leads Claude Fable 5.2 in Science at #1 against #2.
- Which to pick, Coding: As of Nov 22, 2026, Claude Fable 5.2 leads GPT-6 Astra in Coding at #2 against #3.
- Which to pick, Coding › Web dev: As of Nov 22, 2026, Claude Fable 5.2 leads GPT-6 Astra in Coding › Web dev at #1 against #6.
- Which to pick, Long context: As of Nov 22, 2026, Claude Fable 5.2 leads GPT-6 Astra in Long context at #5 against #35.
- Which to pick, Factuality: As of Nov 22, 2026, Claude Fable 5.2 leads GPT-6 Astra in Factuality at #13 against #36.
- Which to pick, Abstraction (ARC): As of Nov 22, 2026, Claude Fable 5.2 and GPT-6 Astra share #1 in Abstraction (ARC).
- Which to pick, Legal: As of Nov 22, 2026, Claude Fable 5.2 leads GPT-6 Astra in Legal at #13 against #21.

As of Nov 22, 2026, Claude Fable 5.2 is tied for state of the art in Coding › Web dev (since Nov 11, 2026; sources LMArena WebDev; methodology v1.0).

As of Nov 22, 2026, GPT-6 Astra is state of the art in Math (since Sep 28, 2026; sources Humanity's Last Exam; methodology v1.0).

## Related comparisons

- [Claude Opus 5.5 vs GPT-6 Astra](https://isitsota.com/compare/claude-opus-5-5-vs-gpt-6-astra)
- [Gemini 4 Carbon vs GPT-6 Astra](https://isitsota.com/compare/gemini-4-carbon-vs-gpt-6-astra)
- [Gemini 4 Argon vs GPT-6 Astra](https://isitsota.com/compare/gemini-4-argon-vs-gpt-6-astra)
- [Claude Fable 5.2 vs Claude Opus 5.5](https://isitsota.com/compare/claude-fable-5-2-vs-claude-opus-5-5)
- [Claude Fable 5.2 vs Gemini 4 Carbon](https://isitsota.com/compare/claude-fable-5-2-vs-gemini-4-carbon)
- [Claude Fable 5.2 vs Gemini 4 Argon](https://isitsota.com/compare/claude-fable-5-2-vs-gemini-4-argon)
- [Gemini 4 Neon vs GPT-6 Astra](https://isitsota.com/compare/gemini-4-neon-vs-gpt-6-astra)
- [Claude Fable 5.2 vs Gemini 4 Neon](https://isitsota.com/compare/claude-fable-5-2-vs-gemini-4-neon)
